* the GNU General Public Licence version 2 or later.
*/
-#include <config.h>
+/*
-#include <cctype>
+If someone desperately needs partial "structures" (such as a few cells of
+an array inset or similar) (s)he could uses the following hack as starting
+point to write some macros:
+
+ \newif\ifcomment
+ \commentfalse
+ \ifcomment
+ \def\makeamptab{\catcode`\&=4\relax}
+ \def\makeampletter{\catcode`\&=11\relax}
+ \def\b{\makeampletter\expandafter\makeamptab\bi}
+ \long\def\bi#1\e{}
+ \else
+ \def\b{}\def\e{}
+ \fi
+
+ ...
+
+ \[\begin{array}{ccc}
+ 1 & 2\b & 3^2\\
+ 4 & 5\e & 6\\
+ 7 & 8 & 9
+ \end{array}\]
+
+*/
+
+
+#include <config.h>
#ifdef __GNUG__
#pragma implementation
#endif
#include "math_parser.h"
-#include "array.h"
#include "math_inset.h"
#include "math_arrayinset.h"
-#include "math_bigopinset.h"
+#include "math_braceinset.h"
+#include "math_casesinset.h"
#include "math_charinset.h"
-#include "math_dotsinset.h"
-#include "math_decorationinset.h"
#include "math_deliminset.h"
-#include "math_fracinset.h"
+#include "math_factory.h"
#include "math_funcinset.h"
-#include "math_funcliminset.h"
+#include "math_kerninset.h"
#include "math_macro.h"
#include "math_macrotable.h"
#include "math_macrotemplate.h"
-#include "math_matrixinset.h"
-#include "math_noglyphinset.h"
+#include "math_hullinset.h"
#include "math_rootinset.h"
-#include "math_scopeinset.h"
#include "math_sqrtinset.h"
#include "math_scriptinset.h"
-#include "math_sizeinset.h"
-#include "math_spaceinset.h"
+#include "math_specialcharinset.h"
+#include "math_splitinset.h"
#include "math_sqrtinset.h"
-#include "math_stackrelinset.h"
-#include "math_symbolinset.h"
-#include "debug.h"
-#include "mathed/support.h"
+#include "math_support.h"
+
#include "lyxlex.h"
+#include "debug.h"
+
#include "support/lstrings.h"
+#include <cctype>
+#include <stack>
+#include <algorithm>
+
using std::istream;
+using std::ostream;
+using std::ios;
using std::endl;
+using std::stack;
+using std::fill;
namespace {
-MathScriptInset * prevScriptInset(MathArray const & array)
+bool stared(string const & s)
{
- MathInset * p = array.back();
- return (p && p->isScriptInset()) ? static_cast<MathScriptInset *>(p) : 0;
+ unsigned const n = s.size();
+ return n && s[n - 1] == '*';
}
-MathInset * lastScriptInset(MathArray & array, bool up, bool down, int limits)
+void add(MathArray & ar, char c, MathTextCodes code)
{
- MathScriptInset * p = prevScriptInset(array);
- if (!p) {
- MathInset * b = array.back();
- if (b && b->isScriptable()) {
- p = new MathScriptInset(up, down, b->clone());
- array.pop_back();
- } else {
- p = new MathScriptInset(up, down);
- }
- array.push_back(p);
- }
- if (up)
- p->up(true);
- if (down)
- p->down(down);
- if (limits)
- p->limits(limits);
- return p;
-}
-
-
-// These are lexical codes, not semantic
-enum lexcode_enum {
- LexNone,
- LexESC,
- LexAlpha,
- LexBOP, // Binary operators or relations
- LexOpen,
- LexClose,
- LexComment,
- LexArgument,
- LexSpace,
- LexNewLine,
- LexOther,
- LexMath,
- LexSelf
+ ar.push_back(MathAtom(new MathCharInset(c, code)));
+}
+
+
+// These are TeX's catcodes
+enum CatCode {
+ catEscape, // 0 backslash
+ catBegin, // 1 {
+ catEnd, // 2 }
+ catMath, // 3 $
+ catAlign, // 4 &
+ catNewline, // 5 ^^M
+ catParameter, // 6 #
+ catSuper, // 7 ^
+ catSub, // 8 _
+ catIgnore, // 9
+ catSpace, // 10 space
+ catLetter, // 11 a-zA-Z
+ catOther, // 12 none of the above
+ catActive, // 13 ~
+ catComment, // 14 %
+ catInvalid // 15 <delete>
};
-lexcode_enum lexcode[256];
+CatCode theCatcode[256];
-const unsigned char LM_TK_OPEN = '{';
-const unsigned char LM_TK_CLOSE = '}';
+inline CatCode catcode(unsigned char c)
+{
+ return theCatcode[c];
+}
+
enum {
- FLAG_BRACE = 1 << 0, // A { needed //}
- FLAG_BRACE_LAST = 1 << 1, // // { Last } ends the parsing process
- FLAG_RIGHT = 1 << 2, // Next right ends the parsing process
- FLAG_END = 1 << 3, // Next end ends the parsing process
- FLAG_BRACK_END = 1 << 5, // // [ Next ] ends the parsing process
- FLAG_AMPERSAND = 1 << 6, // Next & ends the parsing process
- FLAG_NEWLINE = 1 << 7, // Next \\ ends the parsing process
- FLAG_ITEM = 1 << 8, // read a (possibly braced token)
- FLAG_LEAVE = 1 << 9, // marker for leaving the
- FLAG_OPTARG = 1 << 10 // reads an argument in []
+ FLAG_BRACE_LAST = 1 << 1, // last closing brace ends the parsing process
+ FLAG_RIGHT = 1 << 2, // next \\right ends the parsing process
+ FLAG_END = 1 << 3, // next \\end ends the parsing process
+ FLAG_BRACK_END = 1 << 4, // next closing bracket ends the parsing process
+ FLAG_BOX = 1 << 5, // we are in a box
+ FLAG_ITEM = 1 << 7, // read a (possibly braced token)
+ FLAG_BLOCK = 1 << 8, // next block ends the parsing process
+ FLAG_LEAVE = 1 << 9 // leave the loop at the end
};
-struct latex_mathenv_type {
- char const * name;
- char const * basename;
- MathInsetTypes typ;
- bool numbered;
- bool ams;
-};
+void catInit()
+{
+ fill(theCatcode, theCatcode + 256, catOther);
+ fill(theCatcode + 'a', theCatcode + 'z' + 1, catLetter);
+ fill(theCatcode + 'A', theCatcode + 'Z' + 1, catLetter);
+
+ theCatcode['\\'] = catEscape;
+ theCatcode['{'] = catBegin;
+ theCatcode['}'] = catEnd;
+ theCatcode['$'] = catMath;
+ theCatcode['&'] = catAlign;
+ theCatcode['\n'] = catNewline;
+ theCatcode['#'] = catParameter;
+ theCatcode['^'] = catSuper;
+ theCatcode['_'] = catSub;
+ theCatcode['\7f'] = catIgnore;
+ theCatcode[' '] = catSpace;
+ theCatcode['\t'] = catSpace;
+ theCatcode['\r'] = catSpace;
+ theCatcode['~'] = catActive;
+ theCatcode['%'] = catComment;
+}
+
-latex_mathenv_type latex_mathenv[] = {
- {"math", "math", LM_OT_SIMPLE, 0, 0},
- {"equation*", "equation", LM_OT_EQUATION, 0, 0},
- {"equation", "equation", LM_OT_EQUATION, 1, 0},
- {"eqnarray*", "eqnarray", LM_OT_EQNARRAY, 0, 0},
- {"eqnarray", "eqnarray", LM_OT_EQNARRAY, 1, 0},
- {"align*", "align", LM_OT_ALIGN, 0, 1},
- {"align", "align", LM_OT_ALIGN, 1, 1},
- {"alignat*", "alignat", LM_OT_ALIGNAT, 0, 1},
- {"alignat", "alignat", LM_OT_ALIGNAT, 1, 1},
- {"multline*", "multline", LM_OT_MULTLINE, 0, 1},
- {"multline", "multline", LM_OT_MULTLINE, 1, 1},
- {"array", "array", LM_OT_MATRIX, 0, 1}
-};
-int const latex_mathenv_num = sizeof(latex_mathenv)/sizeof(latex_mathenv[0]);
+//
+// Helper class for parsing
+//
+class Token {
+public:
+ ///
+ Token() : cs_(), char_(0), cat_(catIgnore) {}
+ ///
+ Token(char c, CatCode cat) : cs_(), char_(c), cat_(cat) {}
+ ///
+ Token(string const & cs) : cs_(cs), char_(0), cat_(catIgnore) {}
+
+ ///
+ string const & cs() const { return cs_; }
+ ///
+ CatCode cat() const { return cat_; }
+ ///
+ char character() const { return char_; }
+ ///
+ string asString() const;
+ ///
+ bool isCR() const;
+
+private:
+ ///
+ string cs_;
+ ///
+ char char_;
+ ///
+ CatCode cat_;
+};
+bool Token::isCR() const
+{
+ return cs_ == "\\" || cs_ == "cr" || cs_ == "crcr";
+}
-void lexInit()
+string Token::asString() const
{
- for (int i = 0; i <= 255; ++i) {
- if (isdigit(i))
- lexcode[i] = LexOther;
- else if (isspace(i))
- lexcode[i] = LexSpace;
- else
- lexcode[i] = LexAlpha;
- }
-
- lexcode['\t'] = lexcode['\f'] = lexcode[' '] = LexSpace;
- lexcode['\n'] = LexNewLine;
- lexcode['%'] = LexComment;
- lexcode['#'] = LexArgument;
- lexcode['$'] = LexMath;
- lexcode['+'] = lexcode['-'] = lexcode['*'] = lexcode['/']
- = lexcode['<'] = lexcode['>'] = lexcode['='] = LexBOP;
-
- lexcode['('] = lexcode[')'] = lexcode['|'] = lexcode['.'] =
- lexcode['?'] = LexOther;
-
- lexcode['\''] = lexcode['@'] = LexAlpha;
-
- lexcode['['] = lexcode[']'] = lexcode['^'] = lexcode['_'] =
- lexcode['&'] = LexSelf;
-
- lexcode['\\'] = LexESC;
- lexcode['{'] = LexOpen;
- lexcode['}'] = LexClose;
+ return cs_.size() ? cs_ : string(1, char_);
}
+bool operator==(Token const & s, Token const & t)
+{
+ return s.character() == t.character()
+ && s.cat() == t.cat() && s.cs() == t.cs();
+}
+bool operator!=(Token const & s, Token const & t)
+{
+ return !(s == t);
+}
-//
-// Helper class for parsing
-//
+ostream & operator<<(ostream & os, Token const & t)
+{
+ if (t.cs().size())
+ os << "\\" << t.cs();
+ else
+ os << "[" << t.character() << "," << t.cat() << "]";
+ return os;
+}
class Parser {
+
public:
///
- Parser(LyXLex & lex) : is_(lex.getStream()), lineno_(lex.getLineNo()) {}
+ Parser(LyXLex & lex);
///
- Parser(istream & is) : is_(is), lineno_(0) {}
+ Parser(istream & is);
///
- MathMacroTemplate * parse_macro();
+ string parse_macro();
///
- MathMatrixInset * parse_normal();
+ bool parse_normal(MathAtom &);
///
- void parse_into(MathArray & array, unsigned flags);
+ void parse_into(MathArray & array, unsigned flags, MathTextCodes = LM_TC_MIN);
///
int lineno() const { return lineno_; }
+ ///
+ void putback();
private:
///
- int yylex();
- ///
- string lexArg(unsigned char lf, bool accept_spaces = false);
+ string getArg(char lf, char rf);
///
- unsigned char getuchar();
+ char getChar();
///
void error(string const & msg);
///
- void parse_lines(MathGridInset * p, int col, bool numbered, bool outmost);
- ///
- latexkeys const * read_delim();
+ bool parse_lines(MathAtom & t, bool numbered, bool outmost);
private:
///
- istream & is_;
+ void tokenize(istream & is);
///
- int lineno_;
-
+ void tokenize(string const & s);
///
- int ival_;
+ void push_back(Token const & t);
///
- latexkeys const * lval_;
+ void pop_back();
///
- string sval_;
+ Token const & prevToken() const;
+ ///
+ Token const & nextToken() const;
+ ///
+ Token const & getToken();
+ /// skips spaces if any
+ void skipSpaces();
+ /// counts a sequence of hlines
+ int readHLines();
+ ///
+ void lex(string const & s);
+ ///
+ bool good() const;
+ ///
+ int lineno_;
+ ///
+ std::vector<Token> tokens_;
+ ///
+ unsigned pos_;
///
bool curr_num_;
///
};
-unsigned char Parser::getuchar()
+Parser::Parser(LyXLex & lexer)
+ : lineno_(lexer.getLineNo()), pos_(0), curr_num_(false)
{
- char c = 0;
- if (!is_.good())
- lyxerr << "The input stream is not well..." << endl;
- is_.get(c);
- return static_cast<unsigned char>(c);
+ tokenize(lexer.getStream());
+ lexer.eatLine();
}
-string Parser::lexArg(unsigned char lf, bool accept_spaces = false)
+Parser::Parser(istream & is)
+ : lineno_(0), pos_(0), curr_num_(false)
{
- string result;
- unsigned char c = 0;
- while (is_.good()) {
- c = getuchar();
- if (!isspace(c))
- break;
- }
+ tokenize(is);
+}
- if (c != lf) {
- is_.putback(c);
- return result;
- }
-
- unsigned char rg = 0;
- if (lf == '{') rg = '}';
- if (lf == '[') rg = ']';
- if (lf == '(') rg = ')';
- if (!rg) {
- lyxerr[Debug::MATHED] << "Math parse error: unknown bracket '"
- << lf << "'" << endl;
- return result;
+
+void Parser::push_back(Token const & t)
+{
+ tokens_.push_back(t);
+}
+
+
+void Parser::pop_back()
+{
+ tokens_.pop_back();
+}
+
+
+Token const & Parser::prevToken() const
+{
+ static const Token dummy;
+ return pos_ > 0 ? tokens_[pos_ - 1] : dummy;
+}
+
+
+Token const & Parser::nextToken() const
+{
+ static const Token dummy;
+ return good() ? tokens_[pos_] : dummy;
+}
+
+
+Token const & Parser::getToken()
+{
+ static const Token dummy;
+ //lyxerr << "looking at token " << tokens_[pos_] << '\n';
+ return good() ? tokens_[pos_++] : dummy;
+}
+
+
+void Parser::skipSpaces()
+{
+ while (nextToken().cat() == catSpace)
+ getToken();
+}
+
+
+int Parser::readHLines()
+{
+ int num = 0;
+ skipSpaces();
+ while (nextToken().cs() == "hline") {
+ getToken();
+ ++num;
+ skipSpaces();
}
+ return num;
+}
+
+
+void Parser::putback()
+{
+ --pos_;
+}
+
- int depth = 1;
- do {
- unsigned char c = getuchar();
- if (c == lf)
- ++depth;
- if (c == rg)
- --depth;
- if ((!isspace(c) || (c == ' ' && accept_spaces)) && depth > 0)
+bool Parser::good() const
+{
+ return pos_ < tokens_.size();
+}
+
+
+char Parser::getChar()
+{
+ if (!good())
+ lyxerr << "The input stream is not well..." << endl;
+ return tokens_[pos_++].character();
+}
+
+
+string Parser::getArg(char lf, char rg)
+{
+ string result;
+ char c = getChar();
+
+ if (c != lf)
+ putback();
+ else
+ while ((c = getChar()) != rg && good())
result += c;
- } while (depth > 0 && is_.good());
return result;
}
-int Parser::yylex()
+void Parser::tokenize(istream & is)
+{
+ // eat everything up to the next \end_inset or end of stream
+ // and store it in s for further tokenization
+ string s;
+ char c;
+ while (is.get(c)) {
+ s += c;
+ if (s.size() >= 10 && s.substr(s.size() - 10) == "\\end_inset") {
+ s = s.substr(0, s.size() - 10);
+ break;
+ }
+ }
+
+ // tokenize buffer
+ tokenize(s);
+}
+
+
+void Parser::tokenize(string const & buffer)
{
static bool init_done = false;
if (!init_done) {
- lexInit();
+ catInit();
init_done = true;
}
-
- while (is_.good()) {
- unsigned char c = getuchar();
- //lyxerr << "reading byte: '" << c << "' code: " << lexcode[c] << endl;
-
- if (lexcode[c] == LexNewLine) {
- ++lineno_;
- continue;
- } else if (lexcode[c] == LexComment) {
- do {
- c = getuchar();
- } while (c != '\n' && is_.good()); // eat comments
- } else if (lexcode[c] == LexOther) {
- ival_ = c;
- return LM_TK_STR;
- } else if (lexcode[c] == LexAlpha || lexcode[c] == LexSpace) {
- ival_ = c;
- return LM_TK_ALPHA;
- } else if (lexcode[c] == LexBOP) {
- ival_ = c;
- return LM_TK_BOP;
- } else if (lexcode[c] == LexMath) {
- ival_ = 0;
- return LM_TK_MATH;
- } else if (lexcode[c] == LexSelf) {
- return c;
- } else if (lexcode[c] == LexArgument) {
- c = getuchar();
- ival_ = c - '0';
- return LM_TK_ARGUMENT;
- } else if (lexcode[c] == LexOpen) {
- return LM_TK_OPEN;
- } else if (lexcode[c] == LexClose) {
- return LM_TK_CLOSE;
- } else if (lexcode[c] == LexESC) {
- c = getuchar();
- //lyxerr << "reading second byte: '" << c << "' code: " << lexcode[c] << endl;
- string s;
- s += c;
- latexkeys const * l = in_word_set(s);
- if (l) {
- //lyxerr << "found key: " << l << endl;
- //lyxerr << "found key name: " << l->name << endl;
- //lyxerr << "found key token: " << l->token << endl;
- lval_ = l;
- ival_ = l->id;
- return l->token;
- }
- if (lexcode[c] == LexAlpha) {
- sval_.erase();
- while (lexcode[c] == LexAlpha && is_.good()) {
- sval_ += c;
- c = getuchar();
+
+ istringstream is(buffer.c_str(), ios::in | ios::binary);
+
+ char c;
+ while (is.get(c)) {
+
+ switch (catcode(c)) {
+ case catNewline: {
+ ++lineno_;
+ is.get(c);
+ if (catcode(c) == catNewline)
+ ; //push_back(Token("par"));
+ else {
+ push_back(Token(' ', catSpace));
+ is.putback(c);
}
- while (lexcode[c] == LexSpace && is_.good())
- c = getuchar();
- if (lexcode[c] != LexSpace)
- is_.putback(c);
-
- //lyxerr[Debug::MATHED] << "reading: text '" << sval_ << "'\n";
- //lyxerr << "reading: text '" << sval_ << "'\n";
- latexkeys const * l = in_word_set(sval_);
- if (!l)
- return LM_TK_UNDEF;
-
- if (l->token == LM_TK_BEGIN || l->token == LM_TK_END) {
- string name = lexArg('{');
- int i = 0;
- while (i < latex_mathenv_num && name != latex_mathenv[i].name)
- ++i;
- ival_ = i;
- } else if (l->token == LM_TK_SPACE)
- ival_ = l->id;
- else
- lval_ = l;
- return l->token;
+ break;
+ }
+
+ case catComment: {
+ while (is.get(c) && catcode(c) != catNewline)
+ ;
+ ++lineno_;
+ break;
+ }
+
+ case catEscape: {
+ is.get(c);
+ string s(1, c);
+ if (catcode(c) == catLetter) {
+ while (is.get(c) && catcode(c) == catLetter)
+ s += c;
+ if (catcode(c) == catSpace)
+ while (is.get(c) && catcode(c) == catSpace)
+ ;
+ is.putback(c);
+ }
+ push_back(Token(s));
+ break;
}
+
+ default:
+ push_back(Token(c, catcode(c)));
}
}
- return 0;
+
+#if 0
+ lyxerr << "\nTokens: ";
+ for (unsigned i = 0; i < tokens_.size(); ++i)
+ lyxerr << tokens_[i];
+ lyxerr << "\n";
+#endif
}
void Parser::error(string const & msg)
{
lyxerr << "Line ~" << lineno_ << ": Math parse error: " << msg << endl;
+ //exit(1);
}
-void Parser::parse_lines(MathGridInset * p, int col, bool numbered, bool outmost)
-{
+bool Parser::parse_lines(MathAtom & t, bool numbered, bool outmost)
+{
+ MathGridInset * p = t->asGridInset();
+ if (!p) {
+ lyxerr << "error in Parser::parse_lines() 1\n";
+ return false;
+ }
+
+ int const cols = p->ncols();
+
// save global variables
bool const saved_num = curr_num_;
string const saved_label = curr_label_;
+ // read initial hlines
+ p->rowinfo(0).lines_ = readHLines();
+
for (int row = 0; true; ++row) {
// reset global variables
curr_num_ = numbered;
curr_label_.erase();
// reading a row
- int idx = p->nargs() - p->ncols();
- for (int i = 0; i < col - 1; ++i, ++idx)
- parse_into(p->cell(idx), FLAG_AMPERSAND);
- parse_into(p->cell(idx), FLAG_NEWLINE | FLAG_END);
+ for (int col = 0; col < cols; ++col) {
+ //lyxerr << "reading cell " << row << " " << col << "\n";
+ parse_into(p->cell(col + row * cols), FLAG_BLOCK);
+
+ // break if cell is not followed by an ampersand
+ if (nextToken().cat() != catAlign) {
+ //lyxerr << "less cells read than normal in row/col: "
+ // << row << " " << col << "\n";
+ break;
+ }
+
+ // skip the ampersand
+ getToken();
+ }
if (outmost) {
- MathMatrixInset * m = static_cast<MathMatrixInset *>(p);
+ MathHullInset * m = t->asHullInset();
+ if (!m) {
+ lyxerr << "error in Parser::parse_lines() 2\n";
+ return false;
+ }
m->numbered(row, curr_num_);
m->label(row, curr_label_);
if (curr_skip_.size()) {
- m->vskip(LyXLength(curr_skip_), row);
+ m->vcrskip(LyXLength(curr_skip_), row);
curr_skip_.erase();
}
}
-#ifdef WITH_WARNINGS
-#warning Hack!
-#endif
- // no newline
- if (ival_ != -1)
+ // is a \\ coming?
+ if (nextToken().isCR()) {
+ // skip the cr-token
+ getToken();
+
+ // try to read a length
+ //get
+
+ // read hlines for next row
+ p->rowinfo(row + 1).lines_ = readHLines();
+ }
+
+ // we are finished if the next token is an 'end'
+ if (nextToken().cs() == "end") {
+ // skip the end-token
+ getToken();
+ getArg('{','}');
+
+ // leave the 'read a line'-loop
break;
+ }
+ // otherwise, we have to start a new row
p->appendRow();
}
// restore "global" variables
curr_num_ = saved_num;
curr_label_ = saved_label;
+
+ return true;
}
-MathMacroTemplate * Parser::parse_macro()
+string Parser::parse_macro()
{
- if (yylex() != LM_TK_NEWCOMMAND) {
+ string name = "{error}";
+ skipSpaces();
+
+ if (getToken().cs() != "newcommand") {
lyxerr << "\\newcommand expected\n";
- return 0;
+ return name;
+ }
+
+ if (getToken().cat() != catBegin) {
+ lyxerr << "'{' in \\newcommand expected (1)\n";
+ return name;
+ }
+
+ name = getToken().cs();
+
+ if (getToken().cat() != catEnd) {
+ lyxerr << "'}' expected\n";
+ return name;
}
- string name = lexArg('{').substr(1);
- string arg = lexArg('[');
- int narg = arg.empty() ? 0 : atoi(arg.c_str());
- MathMacroTemplate * p = new MathMacroTemplate(name, narg);
- parse_into(p->cell(0), FLAG_BRACE | FLAG_BRACE_LAST);
- return p;
+ string arg = getArg('[', ']');
+ int narg = arg.empty() ? 0 : atoi(arg.c_str());
+
+ if (getToken().cat() != catBegin) {
+ lyxerr << "'{' in \\newcommand expected (2)\n";
+ return name;
+ }
+
+ MathArray ar;
+ parse_into(ar, FLAG_BRACE_LAST);
+ MathMacroTable::create(name, narg, ar);
+ return name;
}
-MathMatrixInset * Parser::parse_normal()
+bool Parser::parse_normal(MathAtom & matrix)
{
- MathMatrixInset * p = 0;
- int t = yylex();
+ skipSpaces();
+ Token const & t = getToken();
- switch (t) {
- case LM_TK_MATH:
- case LM_TK_BEGIN: {
- int i = ival_;
- lyxerr[Debug::MATHED]
- << "reading math environment " << i << " "
- << latex_mathenv[i].name << "\n";
+ if (t.cs() == "(") {
+ matrix = MathAtom(new MathHullInset(LM_OT_SIMPLE));
+ parse_into(matrix->cell(0), 0);
+ return true;
+ }
- MathInsetTypes typ = latex_mathenv[i].typ;
- p = new MathMatrixInset(typ);
+ if (t.cat() == catMath) {
+ Token const & n = getToken();
+ if (n.cat() == catMath) {
+ // TeX's $$...$$ syntax for displayed math
+ matrix = MathAtom(new MathHullInset(LM_OT_EQUATION));
+ MathHullInset * p = matrix->asHullInset();
+ parse_into(p->cell(0), 0);
+ p->numbered(0, curr_num_);
+ p->label(0, curr_label_);
+ } else {
+ // simple $...$ stuff
+ putback();
+ matrix = MathAtom(new MathHullInset(LM_OT_SIMPLE));
+ parse_into(matrix->cell(0), 0);
+ }
+ return true;
+ }
- switch (typ) {
+ if (!t.cs().size()) {
+ lyxerr << "start of math expected, got '" << t << "'\n";
+ return false;
+ }
- case LM_OT_SIMPLE: {
- curr_num_ = latex_mathenv[i].numbered;
- curr_label_.erase();
- parse_into(p->cell(0), 0);
- p->numbered(0, curr_num_);
- p->label(0, curr_label_);
- break;
- }
+ string const & cs = t.cs();
- case LM_OT_EQUATION: {
- curr_num_ = latex_mathenv[i].numbered;
- curr_label_.erase();
- parse_into(p->cell(0), FLAG_END);
- p->numbered(0, curr_num_);
- p->label(0, curr_label_);
- break;
- }
+ if (cs == "[") {
+ curr_num_ = 0;
+ curr_label_.erase();
+ matrix = MathAtom(new MathHullInset(LM_OT_EQUATION));
+ MathHullInset * p = matrix->asHullInset();
+ parse_into(p->cell(0), 0);
+ p->numbered(0, curr_num_);
+ p->label(0, curr_label_);
+ return true;
+ }
- case LM_OT_EQNARRAY: {
- parse_lines(p, 3, latex_mathenv[i].numbered, true);
- break;
- }
+ if (cs != "begin") {
+ lyxerr << "'begin' of un-simple math expected, got '" << cs << "'\n";
+ return false;
+ }
- case LM_OT_ALIGN: {
- p->halign(lexArg('{'));
- parse_lines(p, 2, latex_mathenv[i].numbered, true);
- break;
- }
+ string const name = getArg('{', '}');
- case LM_OT_ALIGNAT: {
- p->halign(lexArg('{'));
- parse_lines(p, 2, latex_mathenv[i].numbered, true);
- break;
- }
+ if (name == "equation" || name == "equation*" || name == "displaymath") {
+ curr_num_ = (name == "equation");
+ curr_label_.erase();
+ matrix = MathAtom(new MathHullInset(LM_OT_EQUATION));
+ MathHullInset * p = matrix->asHullInset();
+ parse_into(p->cell(0), FLAG_END);
+ p->numbered(0, curr_num_);
+ p->label(0, curr_label_);
+ return true;
+ }
- default:
- lyxerr[Debug::MATHED]
- << "1: unknown math environment: " << typ << "\n";
- }
+ if (name == "eqnarray" || name == "eqnarray*") {
+ matrix = MathAtom(new MathHullInset(LM_OT_EQNARRAY));
+ return parse_lines(matrix, !stared(name), true);
+ }
- break;
- }
-
- default:
- lyxerr[Debug::MATHED]
- << "2 unknown math environment: " << t << "\n";
+ if (name == "align" || name == "align*") {
+ matrix = MathAtom(new MathHullInset(LM_OT_ALIGN));
+ return parse_lines(matrix, !stared(name), true);
}
- return p;
-}
+ if (name == "alignat" || name == "alignat*") {
+ int nc = 2 * atoi(getArg('{', '}').c_str());
+ matrix = MathAtom(new MathHullInset(LM_OT_ALIGNAT, nc));
+ return parse_lines(matrix, !stared(name), true);
+ }
+ if (name == "xalignat" || name == "xalignat*") {
+ int nc = 2 * atoi(getArg('{', '}').c_str());
+ matrix = MathAtom(new MathHullInset(LM_OT_XALIGNAT, nc));
+ return parse_lines(matrix, !stared(name), true);
+ }
-latexkeys const * Parser::read_delim()
-{
- int ld = yylex();
- //lyxerr << "found symbol: " << ld << "\n";
- latexkeys const * l = in_word_set(".");
- switch (ld) {
- case LM_TK_SYM:
- case LM_TK_NOGLYPH:
- case LM_TK_SPECIAL:
- case LM_TK_BEGIN: {
- l = lval_;
- //lyxerr << "found key 1: '" << l << "'\n";
- //lyxerr << "found key 1: '" << l->name << "'\n";
- break;
- }
- case ']':
- case '[': {
- string s;
- s += ld;
- l = in_word_set(s);
- //lyxerr << "found key 2: '" << l->name << "'\n";
- break;
- }
- case LM_TK_STR: {
- string s;
- s += ival_;
- l = in_word_set(s);
- //lyxerr << "found key 2: '" << l->name << "'\n";
- }
+ if (name == "xxalignat") {
+ int nc = 2 * atoi(getArg('{', '}').c_str());
+ matrix = MathAtom(new MathHullInset(LM_OT_XXALIGNAT, nc));
+ return parse_lines(matrix, !stared(name), true);
}
- return l;
+
+ if (name == "multline" || name == "multline*") {
+ matrix = MathAtom(new MathHullInset(LM_OT_MULTLINE));
+ return parse_lines(matrix, !stared(name), true);
+ }
+
+ if (name == "gather" || name == "gather*") {
+ matrix = MathAtom(new MathHullInset(LM_OT_GATHER));
+ return parse_lines(matrix, !stared(name), true);
+ }
+
+ lyxerr[Debug::MATHED] << "1: unknown math environment: " << name << "\n";
+ lyxerr << "1: unknown math environment: " << name << "\n";
+ return false;
}
-void Parser::parse_into(MathArray & array, unsigned flags)
+void Parser::parse_into(MathArray & array, unsigned flags, MathTextCodes code)
{
- MathTextCodes yyvarcode = LM_TC_VAR;
-
- int t = yylex();
bool panic = false;
- int brace = 0;
int limits = 0;
- while (t) {
- //lyxerr << "t: " << t << " flags: " << flags << " i: " << ival_
- // << " '" << sval_ << "'\n";
+ while (good()) {
+ Token const & t = getToken();
+
+ //lyxerr << "t: " << t << " flags: " << flags << "'\n";
//array.dump(lyxerr);
//lyxerr << "\n";
if (flags & FLAG_ITEM) {
flags &= ~FLAG_ITEM;
- if (t == LM_TK_OPEN) {
- // skip the brace and regard everything to the next matching
+ if (t.cat() == catBegin) {
+ // skip the brace and collect everything to the next matching
// closing brace
- t = yylex();
- ++brace;
flags |= FLAG_BRACE_LAST;
+ continue;
} else {
- // regard only this single token
+ // handle only this single token, leave the loop if done
flags |= FLAG_LEAVE;
}
}
- if ((flags & FLAG_BRACE) && t != LM_TK_OPEN) {
- error(
- "Expected {. Maybe you forgot to enclose an argument in {}");
- panic = true;
- break;
+ if (flags & FLAG_BLOCK) {
+ if (t.cat() == catAlign || t.isCR() || t.cs() == "end") {
+ putback();
+ return;
+ }
}
- switch (t) {
-
- case LM_TK_ALPHA:
- if (!isspace(ival_) || yyvarcode == LM_TC_TEXTRM)
- array.push_back(new MathCharInset(ival_, yyvarcode));
- break;
-
- case LM_TK_ARGUMENT: {
- MathMacroArgument * p = new MathMacroArgument(ival_);
- //p->code(yyvarcode);
- array.push_back(p);
- break;
+ //
+ // cat codes
+ //
+ if (t.cat() == catMath) {
+ if (flags & FLAG_BOX) {
+ // we are inside an mbox, so opening new math is allowed
+ array.push_back(MathAtom(new MathHullInset(LM_OT_SIMPLE)));
+ parse_into(array.back()->cell(0), 0);
+ } else {
+ // otherwise this is the end of the formula
+ break;
+ }
}
- case LM_TK_SPECIAL:
- array.push_back(new MathCharInset(ival_, LM_TC_SPECIAL));
- break;
+ else if (t.cat() == catLetter)
+ add(array, t.character(), code);
- case LM_TK_STR:
- array.push_back(new MathCharInset(ival_, LM_TC_CONST));
- break;
+ else if (t.cat() == catSpace && code == LM_TC_TEXTRM)
+ add(array, t.character(), code);
- case LM_TK_OPEN:
- array.push_back(new MathScopeInset);
- parse_into(array.back()->cell(0), FLAG_BRACE_LAST);
- break;
+ else if (t.cat() == catParameter) {
+ Token const & n = getToken();
+ array.push_back(MathAtom(new MathMacroArgument(n.character() - '0')));
+ }
- case LM_TK_CLOSE:
- if (flags & FLAG_BRACE_LAST) {
- flags |= FLAG_LEAVE;
+ else if (t.cat() == catBegin) {
+ MathArray ar;
+ parse_into(ar, FLAG_BRACE_LAST);
+#ifndef WITH_WARNINGS
+#warning this might be wrong in general!
+#endif
+ // ignore braces around simple items
+ if (ar.size() == 1 || (ar.size() == 2 && ar.back()->asScriptInset())) {
+ array.push_back(ar);
+ } else {
+ array.push_back(MathAtom(new MathBraceInset));
+ array.back()->cell(0).swap(ar);
}
- break;
-
- case '[':
- array.push_back(new MathCharInset('[', LM_TC_CONST));
- break;
+ }
- case ']':
- if (flags & FLAG_BRACK_END)
- flags |= FLAG_LEAVE;
- else
- array.push_back(new MathCharInset(']', LM_TC_CONST));
- break;
-
- case '^':
- parse_into(
- lastScriptInset(array, true, false, limits)->cell(0), FLAG_ITEM);
- break;
-
- case '_':
- parse_into(
- lastScriptInset(array, false, true, limits)->cell(1), FLAG_ITEM);
- break;
-
- case LM_TK_LIMIT:
- limits = lval_->id;
- //lyxerr << "setting limit to " << limits << "\n";
- break;
-
- case '&':
- if (flags & FLAG_AMPERSAND) {
- flags &= ~FLAG_AMPERSAND;
+ else if (t.cat() == catEnd) {
+ if (flags & FLAG_BRACE_LAST)
return;
- }
- lyxerr[Debug::MATHED]
- << "found tab unexpectedly, array: '" << array << "'\n";
- break;
+ lyxerr << "found '}' unexpectedly, array: '" << array << "'\n";
+ //lyxerr << "found '}' unexpectedly\n";
+ add(array, '}', LM_TC_TEX);
+ }
- case LM_TK_NEWLINE:
- {
- curr_skip_ = lexArg('[');
- if (flags & FLAG_NEWLINE) {
- flags &= ~FLAG_NEWLINE;
- return;
- }
- lyxerr[Debug::MATHED]
- << "found newline unexpectedly, array: '" << array << "'\n";
- break;
+ else if (t.cat() == catAlign) {
+ lyxerr << "found tab unexpectedly, array: '" << array << "'\n";
+ //lyxerr << "found tab unexpectedly\n";
+ add(array, '&', LM_TC_TEX);
}
- case LM_TK_PROTECT:
- break;
-
- case LM_TK_NOGLYPH:
- case LM_TK_NOGLYPHB:
- limits = 0;
- array.push_back(new MathNoglyphInset(lval_));
- break;
-
- case LM_TK_BIGSYM:
+ else if (t.cat() == catSuper || t.cat() == catSub) {
+ bool up = (t.cat() == catSuper);
+ MathScriptInset * p = 0;
+ if (array.size())
+ p = array.back()->asScriptInset();
+ if (!p || p->has(up)) {
+ array.push_back(MathAtom(new MathScriptInset(up)));
+ p = array.back()->asScriptInset();
+ }
+ p->ensure(up);
+ parse_into(p->cell(up), FLAG_ITEM);
+ p->limits(limits);
limits = 0;
- array.push_back(new MathBigopInset(lval_));
- break;
+ }
- case LM_TK_FUNCLIM:
- limits = 0;
- array.push_back(new MathFuncLimInset(lval_));
- break;
+ else if (t.character() == ']' && (flags & FLAG_BRACK_END))
+ return;
- case LM_TK_SYM:
- limits = 0;
- array.push_back(new MathSymbolInset(lval_));
+ else if (t.cat() == catOther)
+ add(array, t.character(), code);
+
+ //
+ // control sequences
+ //
+ else if (t.cs() == "protect")
+ // ignore \\protect, will be re-added during output
+ ;
+
+ else if (t.cs() == "end")
break;
- case LM_TK_BOP:
- array.push_back(new MathCharInset(ival_, LM_TC_BOP));
+ else if (t.cs() == ")")
break;
- case LM_TK_SPACE:
- if (ival_ >= 0)
- array.push_back(new MathSpaceInset(ival_));
+ else if (t.cs() == "]")
break;
- case LM_TK_DOTS:
- array.push_back(new MathDotsInset(lval_));
- break;
-
- case LM_TK_STACK:
- {
- MathStackrelInset * p = new MathStackrelInset;
- parse_into(p->cell(0), FLAG_ITEM);
- parse_into(p->cell(1), FLAG_ITEM);
- array.push_back(p);
- break;
+ else if (t.cs() == "\\") {
+ curr_skip_ = getArg('[', ']');
+ //lyxerr << "found newline unexpectedly, array: '" << array << "'\n";
+ lyxerr << "found newline unexpectedly\n";
+ array.push_back(createMathInset("\\"));
}
+
+ else if (t.cs() == "limits")
+ limits = 1;
+
+ else if (t.cs() == "nolimits")
+ limits = -1;
+
+ else if (t.cs() == "nonumber")
+ curr_num_ = false;
- case LM_TK_FRAC:
- {
- MathFracInset * p = new MathFracInset;
- parse_into(p->cell(0), FLAG_ITEM);
- parse_into(p->cell(1), FLAG_ITEM);
- array.push_back(p);
- break;
- }
+ else if (t.cs() == "number")
+ curr_num_ = true;
- case LM_TK_SQRT:
- {
- unsigned char c = getuchar();
+ else if (t.cs() == "sqrt") {
+ char c = getChar();
if (c == '[') {
- array.push_back(new MathRootInset);
+ array.push_back(MathAtom(new MathRootInset));
parse_into(array.back()->cell(0), FLAG_BRACK_END);
parse_into(array.back()->cell(1), FLAG_ITEM);
} else {
- is_.putback(c);
- array.push_back(new MathSqrtInset);
+ putback();
+ array.push_back(MathAtom(new MathSqrtInset));
parse_into(array.back()->cell(0), FLAG_ITEM);
}
- break;
}
- case LM_TK_LEFT:
- {
- latexkeys const * l = read_delim();
+ else if (t.cs() == "left") {
+ string l = getToken().asString();
MathArray ar;
parse_into(ar, FLAG_RIGHT);
- latexkeys const * r = read_delim();
- MathDelimInset * dl = new MathDelimInset(l, r);
+ string r = getToken().asString();
+ MathAtom dl(new MathDelimInset(l, r));
dl->cell(0) = ar;
array.push_back(dl);
- break;
}
- case LM_TK_RIGHT:
- if (flags & FLAG_RIGHT)
- return;
- error("Unmatched right delimiter");
-// panic = true;
- break;
-
- case LM_TK_FONT:
- {
- MathTextCodes t = static_cast<MathTextCodes>(lval_->id);
- MathArray ar;
- parse_into(ar, FLAG_ITEM);
- for (MathArray::iterator it = ar.begin(); it != ar.end(); ++it)
- (*it)->handleFont(t);
- array.push_back(ar);
- break;
+ else if (t.cs() == "right") {
+ if (!(flags & FLAG_RIGHT)) {
+ //lyxerr << "got so far: '" << array << "'\n";
+ error("Unmatched right delimiter");
+ }
+ return;
}
- case LM_TK_OLDFONT:
- yyvarcode = static_cast<MathTextCodes>(lval_->id);
- break;
-
+/*
case LM_TK_STY:
{
lyxerr[Debug::MATHED] << "LM_TK_STY not implemented\n";
break;
}
- case LM_TK_DECORATION:
- {
- MathDecorationInset * p = new MathDecorationInset(lval_);
- parse_into(p->cell(0), FLAG_ITEM);
- array.push_back(p);
- break;
- }
-
- case LM_TK_NONUM:
- curr_num_ = false;
- break;
+*/
- case LM_TK_FUNC:
- array.push_back(new MathSymbolInset(lval_));
- break;
-
- case LM_TK_UNDEF:
- if (MathMacroTable::hasTemplate(sval_)) {
- MathMacro * m = MathMacroTable::cloneTemplate(sval_);
- for (int i = 0; i < m->nargs(); ++i)
- parse_into(m->cell(i), FLAG_ITEM);
- array.push_back(m);
- m->metrics(LM_ST_TEXT);
- } else
- array.push_back(new MathFuncInset(sval_));
- break;
-
- case LM_TK_MATH:
- case LM_TK_END:
+ else if (t.cs() == "begin") {
+ string const name = getArg('{', '}');
+ if (name == "array") {
+ string const valign = getArg('[', ']') + 'c';
+ string const halign = getArg('{', '}');
+ array.push_back(MathAtom(new MathArrayInset(valign[0], halign)));
+ parse_lines(array.back(), false, false);
+ } else if (name == "split") {
+ array.push_back(MathAtom(new MathSplitInset(1)));
+ parse_lines(array.back(), false, false);
+ } else if (name == "cases") {
+ array.push_back(MathAtom(new MathCasesInset));
+ parse_lines(array.back(), false, false);
+ } else
+ lyxerr << "unknow math inset begin '" << name << "'\n";
+ }
+
+ else if (t.cs() == "kern") {
+#ifdef WITH_WARNINGS
+#warning A hack...
+#endif
+ string s;
+ while (1) {
+ Token const & t = getToken();
+ if (!good()) {
+ putback();
+ break;
+ }
+ s += t.character();
+ if (isValidLength(s))
+ break;
+ }
+ array.push_back(MathAtom(new MathKernInset(s)));
+ }
+
+ else if (t.cs() == "label") {
+ curr_label_ = getArg('{', '}');
+ }
+
+ else if (t.cs() == "choose" || t.cs() == "over" || t.cs() == "atop") {
+ MathAtom p = createMathInset(t.cs());
+ array.swap(p->cell(0));
+ parse_into(p->cell(1), flags, code);
+ array.push_back(p);
return;
+ }
- case LM_TK_BEGIN:
- {
- int i = ival_;
- MathInsetTypes typ = latex_mathenv[i].typ;
-
- if (typ == LM_OT_MATRIX) {
- string const valign = lexArg('[') + 'c';
- string const halign = lexArg('{');
- //lyxerr << "valign: '" << valign << "'\n";
- //lyxerr << "halign: '" << halign << "'\n";
- MathArrayInset * m = new MathArrayInset(halign.size(), 1);
- m->valign(valign[0]);
- m->halign(halign);
-
- parse_lines(m, halign.size(), latex_mathenv[i].numbered, false);
- array.push_back(m);
- //lyxerr << "read matrix " << *m << "\n";
- break;
- } else
- lyxerr[Debug::MATHED] << "unknow math inset " << typ << "\n";
- break;
+/*
+ // Disabled
+ else if (t.cs() == "mbox") {
+ array.push_back(createMathInset(t.cs()));
+ // slurp in the argument of mbox
+
+ MathBoxInset * p = array.back()->asBoxInset();
+ //lyx::assert(p);
}
+*/
- case LM_TK_MACRO:
- array.push_back(MathMacroTable::cloneTemplate(lval_->name));
- break;
-
- case LM_TK_LABEL:
- curr_label_ = lexArg('{', true);
- break;
-
- default:
- error("Unrecognized token");
- lyxerr[Debug::MATHED] << "[" << t << " " << sval_ << "]" << endl;
- break;
+ else if (t.cs().size()) {
+ latexkeys const * l = in_word_set(t.cs());
+ if (l) {
+ if (l->token == LM_TK_FONT) {
+ //lyxerr << "starting font\n";
+ //CatCode catSpaceSave = theCatcode[' '];
+ //if (l->id == LM_TC_TEXTRM) {
+ // // temporarily change catcode
+ // theCatcode[' '] = catLetter;
+ //}
+
+ MathArray ar;
+ parse_into(ar, FLAG_ITEM, static_cast<MathTextCodes>(l->id));
+ array.push_back(ar);
+
+ // undo catcode changes
+ ////theCatcode[' '] = catSpaceSave;
+ //lyxerr << "ending font\n";
+ }
+
+ else if (l->token == LM_TK_OLDFONT) {
+ code = static_cast<MathTextCodes>(l->id);
+ }
+
+ else if (l->token == LM_TK_BOX) {
+ MathAtom p = createMathInset(t.cs());
+ parse_into(p->cell(0), FLAG_ITEM | FLAG_BOX, LM_TC_BOX);
+ array.push_back(p);
+ }
+
+ else {
+ MathAtom p = createMathInset(t.cs());
+ for (MathInset::idx_type i = 0; i < p->nargs(); ++i)
+ parse_into(p->cell(i), FLAG_ITEM);
+ array.push_back(p);
+ }
+ }
+
+ else {
+ MathAtom p = createMathInset(t.cs());
+ for (MathInset::idx_type i = 0; i < p->nargs(); ++i)
+ parse_into(p->cell(i), FLAG_ITEM);
+ array.push_back(p);
+ }
+ }
- } // end of big switch
if (flags & FLAG_LEAVE) {
flags &= ~FLAG_LEAVE;
break;
}
+ }
- if (panic) {
- lyxerr << " Math Panic, expect problems!" << endl;
- // Search for the end command.
- do {
- t = yylex();
- } while (is_.good() && t != LM_TK_END && t);
- } else {
- t = yylex();
- }
+ if (panic) {
+ lyxerr << " Math Panic, expect problems!\n";
+ // Search for the end command.
+ Token t;
+ do {
+ t = getToken();
+ } while (good() && t.cs() != "end");
}
}
-void parse_end(LyXLex & lex, int lineno)
-{
- // Update line number
- lex.setLineNo(lineno);
-
- // reading of end_inset
- while (lex.isOK()) {
- lex.nextToken();
- if (lex.getString() == "\\end_inset")
- break;
- lyxerr[Debug::MATHED] << "InsetFormula::Read: Garbage before \\end_inset,"
- " or missing \\end_inset!" << endl;
- }
-}
} // anonymous namespace
-
-MathArray mathed_parse_cell(string const & str)
+void mathed_parse_cell(MathArray & ar, string const & str)
{
istringstream is(str.c_str());
- Parser parser(is);
- MathArray ar;
- parser.parse_into(ar, 0);
- return ar;
+ mathed_parse_cell(ar, is);
+}
+
+
+void mathed_parse_cell(MathArray & ar, istream & is)
+{
+ Parser(is).parse_into(ar, 0);
}
-MathMacroTemplate * mathed_parse_macro(string const & str)
+string mathed_parse_macro(string const & str)
{
istringstream is(str.c_str());
Parser parser(is);
return parser.parse_macro();
}
-MathMacroTemplate * mathed_parse_macro(istream & is)
+string mathed_parse_macro(istream & is)
{
Parser parser(is);
return parser.parse_macro();
}
-MathMacroTemplate * mathed_parse_macro(LyXLex & lex)
+string mathed_parse_macro(LyXLex & lex)
{
Parser parser(lex);
- MathMacroTemplate * p = parser.parse_macro();
- parse_end(lex, parser.lineno());
- return p;
+ return parser.parse_macro();
}
-MathMatrixInset * mathed_parse_normal(string const & str)
+bool mathed_parse_normal(MathAtom & t, string const & str)
{
istringstream is(str.c_str());
Parser parser(is);
- return parser.parse_normal();
+ return parser.parse_normal(t);
}
-MathMatrixInset * mathed_parse_normal(istream & is)
+bool mathed_parse_normal(MathAtom & t, istream & is)
{
Parser parser(is);
- return parser.parse_normal();
+ return parser.parse_normal(t);
}
-MathMatrixInset * mathed_parse_normal(LyXLex & lex)
+bool mathed_parse_normal(MathAtom & t, LyXLex & lex)
{
Parser parser(lex);
- MathMatrixInset * p = parser.parse_normal();
- parse_end(lex, parser.lineno());
- return p;
+ return parser.parse_normal(t);
}
-